Vulnerability in Vllm
CVE-2026-44223
vLLM is an inference and serving engine for large language models (LLMs). From 0.18.0 to before 0.20.0, the extract_hidden_states speculative decoding proposer in vLLM returns a tensor with an incorrect shape after the first decode step, causing a RuntimeError that crashes the EngineCore process. The crash is triggered when any request in the batch uses sampling penalty parameters (repetition_penalty, frequency_penalty, or presence_penalty). A single request with a penalty parameter (e.g., "repetition_penalty": 1.1) is sufficient to crash the server. This vulnerability is fixed in 0.20.0.
Published · last modified .
CVSS v3 metric
CVSS v3 base score 6.5 (Medium). Vector: CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:N/I:N/A:H.
EPSS exploit prediction
EPSS 0.004, 31.0th percentile (scored ); unchanged over 30 days. About EPSS.
Affected products
- Vllm
- Vllm-Project Vllm — versions >= 0.18.0, < 0.20.0
Weakness classification (CWE)
References
- github.com/vllm-project/vllm/pull/38610 (Patch, Issue Tracking)
- github.com/vllm-project/vllm/security/advisories/GHSA-83vm-p52w-f9pw (Mitigation, Vendor Advisory)